A study on electrode fabrication and operation variables affecting the performance of anion exchange membrane water electrolysis

Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec

A study on electrode fabrication and operation variables affecting the performance of anion exchange membrane water electrolysis Ahyoun Lima,b , Hyoung-juhn Kima,c, Dirk Henkensmeiera,c,d, Sung Jong Yooa,c, Jin Young Kima,c,d , So Young Leea , Yung-Eun Sungb , Jong Hyun Janga,c,d,* , Hyun S. Parka,c,* a

Center for Hydrogen Fuel Cell Research, Korea Institute of Science and Technology (KIST), Seoul 02792, Republic of Korea School of Chemical and Biological Engineering, Seoul National University, Seoul 08826, Republic of Korea Division of Energy & Environment Technology, KIST School, University of Science and Technology (UST), Seoul 02792, Republic of Korea d Green School, Korea University, Seoul 02841, Republic of Korea b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 26 December 2018 Received in revised form 1 April 2019 Accepted 3 April 2019 Available online 11 April 2019

Polymer electrolyte membrane water electrolysis has been proposed to address production of high purity hydrogen for storage of excess renewable energy. Among them, alkaline electrolyte membrane based water electrolysis (AEMWE) has an advantage in the aspect of material costs, e.g. from non-noble catalysts and membrane, but suffers from lower performance compared to proton exchange membrane based water electrolysis (PEMWE). However, there are fewer researches on single cell MEA and operation study compared to material research to enhance AEMWE performance. Here, we analyze the effect of the cell construction and operation factors, i.e MEA pressing, torque of cell assembly, electrolyte pre-feed methods, and operation temperature, to obtain high performance in AEMWE single cell operation. 97.5 % current improves at 1.8 V by applying optimized torque. 94 % decrease of ohmic resistance are achieved from electrolyte pre-feeding. 50 mA cm 2 of current density is enhanced at 0.591 V of overvoltage per 10  C temperature increase due to higher ionic conductivity and reaction kinetics. These factors significantly affect internal factors such as not only material property during operation but also, catalysts structure and contact in MEA, leading 4.3 times progress of current density from 0.242 to 1.045 A cm 2 at 1.8 Vcell. © 2019 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

Keywords: Anion exchange membrane water electrolysis membrane electrode assembly electrolyzer operation electrocatalysis

Introduction The impacts of global warming and climate change due to increased greenhouse gas emissions have become a serious issue in recent decades [1,2]. In order to mitigate the effects of climate change, environmentally clean and economically efficient noncarbon fuel should be produced and used to reduce the carbon emissions from fossil fuel consumption [3]. Hydrogen is considered an ideal energy storage medium, converting energy from renewable power sources into chemical bonds with a high energy density, i.e. 33.3 kW h kg 1 for pressurized hydrogen (200 bar) or for liquid hydrogen ( 253  C) [4,5]. The use of hydrogen in fuel cell technology is mature and has already been commercialized for large stationary and mobile applications [6]. However, environmentally sustainable production of hydrogen should also be implemented to form complete hydrogen-based energy cycles.

* Corresponding authors at: Center for Hydrogen Fuel Cell Research, Korea Institute of Science and Technology (KIST), Seoul 02792, Republic of Korea E-mail addresses: [email protected] (J.H. Jang), [email protected] (H.S. Park).

Hydrogen can be generated by different technologies including hydrocarbon steam reforming, gas-shift reaction, biomass gasification, and water splitting through pyrolysis, plasma reforming, and electrolysis. Steam reforming of natural gas by a gas-shift reaction is the most commonly used industrial process for largescale hydrogen production due to its high thermal efficiency (>85%) and low production cost [7]. However, it generates carbon monoxide or dioxide as byproducts from the natural gas decomposition. Biomass gasification can play a significant role in the steady production of hydrogen; however, the noncombustion process still produces a low level of carbon dioxide [8]. Water electrolysis has zero carbon footprint if the process is operated by renewable power sources, e.g. wind, geothermal, or solar energy, opening the door to the possibility of carbon-free and eco-friendly energy cycles. Water electrolysis is categorized into different methods based on the type of electrolyte, operation temperature, and pressure [9]. For example, alkaline water electrolysis (AWE) using diaphragm separators is operated in basic aqueous solutions at low temperature (<80  C). It is a mature technology and has been industrially implemented for hydrogen production more than 50 years.

https://doi.org/10.1016/j.jiec.2019.04.007 1226-086X/© 2019 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

However, AWE operates in low current densities compared to those using polymer electrolyte membranes [10]. Water electrolysis based on a polymer electrolyte membrane makes the process more energy-efficient and easy to operate for pressurized hydrogen production compared to that using liquid electrolytes, where the membrane separates the hydrogen and oxygen evolution reactors [11]. There are two types of polymer-based water electrolysis; proton exchange membrane based water electrolysis (PEMWE) and anion exchange membrane based water electrolysis (AEMWE). PEMWE has high operational efficiency from the high ionic conductivity (100  20 mS cm 1), and high current density above 2 A cm 2, and high operation pressure up to 700 bar on the prototype level [12]. However, the materials used in PEMWE must sustain its property in highly corrosive condition, i.e. low pH, leading cost issue of using noble catalysts such as Ir and Ru [13]. By contrast, AEMWE alleviates the cost problem as inexpensive hydrocarbon-based membrane and non-noble catalysts can be employed in the alkaline medium. However, AEMWE generally presents low performance compared to PEMWE due to low thermal stability of anion conducting membranes and low activities of non-noble catalysts. In other words, AEMWE has much room to improve the performance. The basic principles of polymer membrane based water electrolysis are described below. Cathode half-reaction: 4 H2O + 4 e → 2 H2 + 4 OH 0.828 VSHE)

Anode half-reaction: O2 + 2 H2O + 4 e → 4 OH VSHE)

(pH 14, E0 = (1)

(pH 14, E0 = 0.401 (2)

A schematic structure of the membrane reactor comprising the cathode and anode on both sides of the ion-conducting membrane

411

is shown in Fig. 1, where the proton reduction and water oxidation reactions occur at the cathode and anode, respectively (Eqs. 1 and 2). In the device operation, a basic aqueous solution is provided to the anode, with a voltage larger than 1.23 V applied between the two electrodes. At the anode, water is oxidized to produce oxygen molecules and electrons. Ionic charge balance of the two electrodes is maintained by hydroxide ion diffusion through the polymer membrane while the proton is reduced at the cathode. Efficient electrochemical reactions require that the kinetics at the catalyst surfaces, ionic transport through the membrane and catalyst layers, and the mass transport of water and gases through the diffusion and catalyst layers be optimized [11,14]. Inexpensive and efficient device components for alkaline water electrolysis have been widely studied for decades with the goal of improved hydrogen and oxygen evolution reaction (HER and OER) kinetics [15]. In particular, extensive research has been conducted on the development of non-noble catalysts and ion-conducting membranes in alkaline conditions. For non-noble catalysts for HER in alkaline conditions, Ni-based alloy metals or metal oxides, e.g. Ni–Mo, Ni–Zn, Ni–Co, NiFeOx, and Ni-based composites have been rigorously studied, showing promising electrocatalytic activity and stability [16,17]. As examples of promising OER non-noble catalysts, metal oxides such as MnOx, CoOx, Cu0.81Co2.19O4 and NiOx [18–21], FeOOH [22] and NiFe-OS [23] are reported, e.g. 10 mA cm 2 at overpotentials of less than 0.5 V in alkali media. In addition to the catalysts, alkaline-exchange membranes have also been developed, demonstrating high anionic conductivity ranging from 7 to 100 mS cm 1, which is comparable to the proton conductivity of Nafion [24–32]. However, the state of operation performance of AEMWE is still significantly lower than that in PEMWE. Despite the large amounts of catalysts used with the developed anionconducting polymer electrolytes, AEMWE is often reported to

Fig. 1. Schematic procedure of fabrication and operation of the anion exchange membrane water electrolysis cell.

412

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

have performances less than 20% of those of PEMWE [14,24,33]. One of the reasons for low performance is the instability of AEM at high temperature and pressure compared to PEM (e.g. Nafion), which obstructs to scale up AEMWE to commercial level [28]. In addition to material research, determining the proper electrode structures and operation factors of AEMWE to obtain high electrolysis efficiency is nontrivial [10,11,18,29,34]. To understand the electrochemical behavior of AEMWE in conjunction of complex mass and ionic transport, mathematical modeling was employed to describe experimental results including exchange current density, membrane thickness and the liquid saturation [35]. Many aspects of the electrode structures, e.g. catalyst layer porosity, polymer binder amount and distribution, and exposed active surface area of catalyst, and operation factors, e.g. electrolyte acidity, solution feed methods, or temperature, should be optimized in the device fabrication and operation for efficient production of hydrogen by water splitting hydrogen. For example, the content of a non-ionomeric polytetrafluoroethylene (PTFE) binder in the anode catalyst layer was studied for the efficient and stable operation of AEMWE. The amount and distribution of PTFE in the catalyst layer is one of the essential factors determining the catalyst layer structure. Cell assembly methods, e.g. hot press conditions, were investigated to improve the performance in AEMWE [14]. Hot press conditions (pressure, temperature, process duration) in the electrode fabrication process can alter the electrode structure and pore distribution of the catalyst layer so to change the exposed catalyst area and mass transfer resistances. Further, a few studies reported the effects of device operation factors on the performance of AEMWE, including temperature [36], pH and concentration of the electrolyte solution [24,37], and electrolyte supply method [11]. The significant dependency of electrolysis performance on various electrode fabrication and operation factors shown in several previous reports indicates that the device performance is determined by complex ensemble effects of many different aspects of cell construction and operation. Therefore, concurrent optimization of different process components from electrode fabrication to device operation is important to understand the effect of each item on the resulting AEMWE performance. In this work, we focused to study the influence of cell fabrication and operation parameters with the standard catalysts, iridium oxide and platinum at anode and cathode. We selected four factors, i.e. fabrication pressing conditions for (i) electrode and (ii) cell assembly; (iii) electrolyte pre-feed methods; and (iv) operation temperatures, to clarify the different effect of each items on the AEMWE performance. By the combinatorial controls of the

fabrication and operation variable of AEMWE, the impact of each factor on device performance is revealed, resulting in significant improvement of the AEMWE device activity. Experimental Electrode preparation The membrane electrode assembly (MEA) was fabricated by the catalyst-coated substrate (CCS) method. Iridium oxide (IrO2, Premion1, Alfa Aesar) and platinum on carbon (Pt/C, Pt 46.5 wt. %, Tanaka K.K) were used as OER and HER catalysts, respectively. In this study, a metal oxide and precious metal were used, since they are the standard water splitting catalysts, while other electrode fabrication and operational factors were investigated. Carbon paper (TGP-H-120, Toray) and Titanium paper (250 mm, Bekaert) were utilized as the conducting substrates supporting the catalysts for the cathode and anode, respectively. The catalyst ink suspension containing catalyst powder, PTFE (60 wt.% PTFE dispersion in H2O, Aldrich), distilled water, and isopropyl alcohol was sprayed onto the GDL and Ti paper to form the catalyst layer. Before spraying, the catalyst suspension was homogenized in an ultrasonic bath for 1 h. The content of PTFE binder was 20 and 9.1 wt.% for the anode and cathode catalyst layers, respectively. Then, the electrodes were sintered at 350  C in Ar gas. Finally, an anion exchange membrane (FAA-3-PK-75) was sandwiched, or hot-pressed, between two sintered electrodes to form the MEA of 6.25 cm2 in the electrolytic cell. Hot pressing was performed at 50  C and 395 psi. The temperature and pressure were not controlled to higher value because of the degradation problem of anion exchange membrane. For the electrochemical measurement, Au-coated Ti and graphite bipolar plates were used as the current collectors on the anode and cathode sides of the cell, respectively. The assembly torque used to construct the cell was controlled as shown in Table 1. A pressure measurement film (Prescale LLLW, Fujifilm) was utilized to measure the contact pressure applied between the electrode and current collector under different cell assembly torques. Electrochemistry and physical characterization The electrolysis performance of AEMWE was measured using a high-current potentiostat (HCP-803, Bio-Logic). The device was operated at 50  C unless otherwise stated. The current and voltage relationship of the AEMWE was obtained during the voltage cycling, which ranged from 1.5 to 2.2 Vcell at a scan rate of 20 mV

Table 1 AEMWE current density at 1.8 V with different variables employed in the cell fabrication and operation. The electrolysis current was obtained at 50th cycle in voltage cycle operation from 1.5 to 2.2 V except the measurements at different operation temperatures. For the temperature controlled operations, current density at 1.8 V was averaged during 60 min of chronoamperometry. Sample #

Controlled variable

MEA press time (min)

Cell assemble torque (Nm)

Pre-feed time (h)

Operation temp. ( C)

j at 1.8 V (mA/cm2)

1 2 3 4 5 6 7 8 9 10 11 12 13

Fabrication

0 1 3 0

4

24

50

24

50

0 24 24

50

478 457 293 242 478 356 303 478 460 607 625 870 983

MEA press time Cell assemble torque

Operation

Pre-feed time

0

3 4 5 4

Temp.

0

4

50 60 70 80 90

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

s 1. Before the water splitting operation, the electrolyte solution was pre-fed to the anode for 24 h, if necessary (Table 1). During electrolysis, 0.5 M KOH aqueous electrolyte solution was fed to the anode with a 1 mL min 1 flow rate. In between electrolysis operations, electrochemical impedance spectroscopy (EIS) was conducted at 1.8 Vcell with an AC frequency ranging from 10 kHz to 30 mHz and an alternating voltage amplitude of 10 mV. The surface morphology of the electrodes was observed by scanning electron microscopy (SEM, Teneo VSTM). The porosity of the catalyst layer coated on the gas diffusion layer was studied by mercury porosimetry (Mercury intrusion porosimetry, Autopore 9600). Result and discussion Effects of the electrode fabrication and operation factors on the hydrogen production performance by AEMWE were studied and are summarized in Table 1 and Fig. 2. Three main parameters—the apparent electron transfer kinetics, mass transport, and ion conductivity—considerably determine the performance of electrolytic devices. However, the key factors of AEMWE mentioned are greatly affected by the electrode fabrication methods and operation variables as shown below. For example, the apparent electron transfer rate of electrocatalysis is directly proportional to the electrochemically active surface area (ECSA) of the catalyst layer [38]. The ion conductivity and chemical mass transport including the ECSA are also determined by the catalyst layer structure including the distribution of ionomeric binder and pores in the catalyst layer [14,33,39]. The ionic conductivity of the polymeric binder and membrane electrolyte is strongly influenced by humidity or chemical environments constructed under specific operation conditions. As shown in Table 1 and Fig. 2, the AEMWE performance varied by more than three times from 293 to 983 mA cm 2 at approximately 1.8 Vcell when conducting combinatorial control of the catalyst layer and MEA construction methods and operation variables. The effects of each variable on the cell performance are discussed below in detail. Effect of membrane electrode and device assembly method Previous reports of single-cell tests of AEMWE have shown MEAs fabricated both using hot pressing [14,40–42] and without heating or pressing [11,43–47]. In the previous study investigating

413

the MEA pressing process, hot pressing (395 psi of applied pressure at 25, 50, and 80 0C for 1 min) was performed to enhance the performance by reducing the interface resistance between the GDL and the membrane [14]. Cho et al. reported that the higher pressing temperature has a positive effect on polarization resistance, but ohmic resistance is increased due to membrane degradation if the process is conducted above a certain temperature; the optimum hot press temperature was 50  C [13]. In this study, the pressing time was controlled (0, 1, and 3 min) at 50  C and 395 psi. The highest water electrolysis performance was shown in the case of MEA without pressing (Fig. 3(a)), which differed from previous studies where the press process was advantageous for the water splitting, probably due to different membrane properties, i.e., the thickness, mechanical strength, etc., employed in MEA (see below). In this work, the MEAs were constructed either without or with 1 or 3 min of hot pressing (see Experimental section and Table 1). In the lower cell voltage region, the current densities of MEAs formed by 1 and 3 min pressing were similar but were lower than the current density of the MEA constructed without pressing. For example, the electrolysis current densities were 118, 118, and 161 mA cm 2 at 1.6 Vcell for the cell fabricated with 3, 1, and 0 min of pressing, respectively. The decreased performance of MEA assembled with hot pressing can be explained by the reduced active surface area of the catalysts by increase PTFE coverage of catalytic active sites and structural deformation of the catalyst layers after the press process (Fig. 3). The electrolysis performance at low overpotentials, or low current densities, is largely determined by the reaction kinetics and catalyst activity rather than the ionic conductivity or mass transport. In other words, the ohmic drop (iR) is smaller in the low current region, and mass transport hindrance is also negligible when the reaction rate is low. The water splitting current in MEAs with 1 and 3 min hot pressing was identical at the voltages smaller than 1.7 Vcell, but the MEA with 1 min pressing shows an approximately 20% higher electrolysis current compared to that with 3 min pressing. As the press time increases, pores in catalysts becomes smaller by agglomeration of PTFE surrounded catalysts (Fig. 3(c)). In addition, the binder, i.e. PTFE, is hydrophobic and non-ion conducting. It induces harsher condition for reactant water and ion approaching to catalysts. The mass limitation or ohmic voltage drop issue becomes more serious in high current operation. IV curve also reveals that larger current differences at a

Fig. 2. Performance summary of AEMWE cells with different fabrication and operation variables. The displayed performance is the current density at 1.8 V measured at 50th cycle in voltage cycle operation from 1.5 to 2.2 V. However, for the temperature controlled operation, current density at 1.8 V was averaged during 60 min of chronoamperometry. Factors of the AEMWE fabrication and operation are summarized in Table 1.

414

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

Fig. 3. Current-voltage relationship (a), Nyquist plot (b), and SEM images of the anode catalyst layer surface (c) of AEMWE cells with different press time used in the electrode fabrication (0, 1, and 3 min.). EIS was measured at 500th cycle at 1.8 V.

higher cell voltage, which implies a longer pressing time leads to increased mass transport resistance in addition to reduced catalyst surfaces. The electrochemical and structural impacts of the hot press process on the catalyst layer are revealed in EIS plots and SEM images (Fig. 3(b) and (c)). The high frequency x-axis intercept of the Nyquist plot in EIS spectra indicates that the hot press process induces an increase in ohmic resistance of approximately 0.25 and

0.31 V cm2 without and with the hot press process, respectively (Fig. 3(b)). The hot press process also noticeably changed the porous composite of catalyst and polymeric binder seen at the catalyst layer surface (porosity shown in Fig. 4) and the PTFE binder penetrated the interparticle spaces and blocked the anode catalyst surface (Fig. 3(c)). However, the effect of pressing was not clear in the cathode surfaces under SEM measurements (Figure S1). As the anode and cathode catalyst are layered on metallic supports and

Fig. 4. Primary pore (a) and secondary pore (b) distribution of cathode with different press time used in the electrode fabrication (0, 1, and 3 min.). The pore distribution is measured by mercury porosimetry.

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

carbon with different mechanical properties, e.g. structural rigidity, the press process results in the increased ohmic, charge, and mass transport resistances of each electrode differently. The structural deformation of the catalyst layer by hot pressing is the main reason behind the performance degradation as no change in membrane conductivity and thickness was observed after the pressing process (see below). To further understand the effect of pressing on the structural deformation of MEA, the pore volume distribution of the electrodes was measured through mercury porosimetry (Fig. 4). The primary pores between the catalysts and polymer composites in the electrodes were small, with a diameter of a few tens to hundreds of nanometers (Fig. 4(a)). The peak pore volume of the cathode catalyst layer decreased from approximately 0.18 to 0.13 mL g 1 at a pore diameter of 0.038 to 0.034 mm in the cathodes after the hot press. However, the effect of hot press duration was negligible in the primary pore distribution and no meaningful difference was detected between cathodes with 1 and 3 min pressing. The reduced primary pore volume and distribution changes in the porosimetry clearly demonstrate that the press process decreases the electrocatalytic active surface area and causes severe electrode agglomeration. As shown in Fig. 4(b), the peak volume of the larger secondary pores of the cathode also decreased by approximately 30 % from 5 to 3.5 mL g 1 due to the press process. Along with the volume changes, the press process decreases the peak secondary pore diameter from 27.58 to 25.62 mm. The secondary pores exist mainly in the GDL since the

415

catalyst layer is thinner (1.96 mm) than the secondary pore size, so the changes in the secondary pores imply that deformation occurred in the GDL and impacted the mass transport of produced H2, O2 and water through the diffusion medium. Similarly, the pore structure of the anode was clearly deformed by the hot press process, as shown in the supporting information (Figure S1 and S2). However, the weight-averaged pore volume of the anode coated on the titanium GDL is an order smaller than that of those on the carbon GDL. In addition to the electrode fabrication method, one of the assembly variables in the device construction step—the torque applied to the cell assembly—was controlled which affects the electric contact resistance of the MEA. Using carbon and titanium GDLs with thicknesses of 370 and 250 mm, respectively, the surface pressure applied between the GDL and current collector was measured at different cell assembly torques. As water electrolysis forms liquid and gaseous reactants and products, the device must be airtight to prevent fluid leakage. The contact between the electrode and current collector should also be intact and allow electric conductivity through the end-plate to the catalyst surface. However, when the assembly torque is too high, the cell components, specifically the catalyst layer, GDL, and membrane electrolyte, would be mechanically damaged between the two metal end-plates. Any non-ideal deformation of the cell components induces performance degradation because it hinders the electronic, ionic, and mass transportation through the devices. For the device components used in this study, we found that 3 Nm of

Fig. 5. (a) Ohmic resistance change in the continuous solution feed of 0.5 M KOH electrolyte solution to AEMWE single cell. CVs at 50th (b) and 500th cycle (c) with and without 24 h of electrolyte pre-feed to AEMWE cell. The ohmic resistance was measured in EIS at open circuit voltage. CVs were measured at a scan rate of 20 mV/s.

416

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

cell assembly torque was not high enough to prevent fluid leakage through the gaps between the electrodes and gaskets (Figure S3(a) and (b)). When the assembly torque increased from 3 to 4 Nm and 5 Nm, a uniform pressure applied to the electrode was observed using the pressure measurement film (Figure S3(c) and (d)). The water splitting performance was also improved with the increased assembly torque (Figure S4). However, if the assembly torque was too large, e.g. 5 Nm, the performance was decreased compared to the case of 4 Nm presumably by the mechanical deformation that occurred under the high pressure. Thus, we found the optimized torque of cell assembly to be 4 Nm in this study, which resulted in an electrolysis current density of 478 mA cm 2 at 1.8 Vcell (Figure S4). Effect of electrolyte solution feed method and operation temperature The membrane in the AEMWE is the core component, separating the gas product and transporting ions to complete the electrochemical reaction of the cell. High ionic conductivity of the membrane is essential for high performance of water electrolysis with low ohmic loss. The hydroxide ion conduction in the AEM are mainly through the Grotthuss mechanism as in water [48,49]. In the Grotthuss mechanism, hydroxide ions are diffused via hydrogen bonded water molecules. Thus, ion diffusion occurs along with the water in the membrane and the ionic conductivity of the electrolyte membrane is a significant function of the water content. In order to ensure facile ion transport through the membrane from the beginning of electrolysis operation, the electrolyte membrane should be sufficiently wetted. The polymer functional group in the membrane should possess OH rather than other anions such as Br , since ion conduction in the membranes significantly relies not only on the mobility but also on the concentration of hydroxide ions. This anion substitution process is called OH doping in the anion exchange electrolyte membrane. Luo et al. reported that ion conductivity in an anion exchange membrane is significantly affected by alkali doping, which decreases the conductivity activation energy with increased doping until reaching saturation [50]. In this study, the ohmic resistance, i.e. the ion transfer resistance, was measured at an open-circuit voltage by EIS. The ohmic resistances showed a sharp decrease in the first 5 h of the OH- doping process using a 0.5 M KOH solution with a flow rate of 1 mL min 1 (Fig. 5(a)). The decrease in ohmic resistance was

smaller after the 5 h pre-feeding process, then decreased more gradually for up to approximately 24 h. The water splitting performance measured after different prefeeding durations indicates that the MEA activation, or OHdoping, requires a longer time to reach a level of doping that ensures its maximum current (Figure S5). The results show that the MEA requires additional time to stabilize, as demonstrated in the comparison between the case of 8 and 24 h feeding times (Figure S5). The initial electrolysis performance was higher and showed a lower ohmic resistance with 24 h of pre-feeding compared to those without or with only 8 h of pre-feeding. The activation step is related to the fluid flow through the catalyst layer, and insignificant performance enhancement or ohmic resistance drop occurs if the AEM is simply immersed in a 1 M KOH aqueous solution for 24 h (Figure S5). Since the ionic transport limitation becomes important given the large current flows, the performance differences become larger with the applied cell voltage increase between the devices with different pre-feeding methods. The electrolysis current density differences were 53 and 361 mA cm 2 at 1.6 and 2.2 Vcell respectively, between MEA with 0 and 24 h of pre-feeding operation at the beginning of the cell operation. However, the performance difference caused by the pre-feeding process disappears when the electrolysis cell is operated for a few hours (Fig. 5(b) and 5(c)). It is highly likely that electrolysis does what the pre-soak is doing in the initial AEMWE operation. The gradual performance enhancement of AEMWE in continuous operation without the pre-feed process implies that the catalyst layer and electrolyte membrane is doped by OH- during the water splitting operation and the pre-feeding is only effective at the initial stage of AEMWE cell operation (Fig. 5(c)). Lastly, the effect of temperature was investigated by controlling the cell operation temperature from 50 to 90  C. The enhancement of electrolysis performance with temperature increase was also briefly reported elsewhere. [36,51] In general, the cell voltage considering different voltage loss terms is expressed by Eq. (3), ECell = E + hCa + hAn. + iROhm + iR’

(3) 

where the open-circuit voltage (E ) of water electrolysis is the thermodynamic value and a function of the temperature (Figure S6). The variation of E can be described by Eq. (4), where (d E /dT) is the temperature coefficient of the electrochemical reactions. E (T2) = E (T1) + (d E /dT)(T2-T1)

(4)

Fig. 6. Chronoamperograms (a) and EIS Nyquist plots (b) of the AEMWE cell operated at different temperatures from 50 to 90  C. The chronoamperometry and EIS was performed at the identical overvoltage of 0.59 V at different operational temperatures.

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

In this equation, the isothermal temperature coefficient, (d E / dT), for the OER reaction is calculated as 0.846 mV deg 1 considering the entropy changes near room temperature given that HER is 0.000 mV deg 1 [52]. Thus, chronoamperometry was used to consider the onset of potential changes in water electrolysis at different temperatures (Fig. 6(a)). The overvoltage applied to the cell at each temperature is the same, 0.591 V, and the increase in current density was approximately 50 mA cm 2 per 10  C. This result implies that the ionic conductivity (Figure S7) and catalytic activity increase at higher temperatures. In addition to the improved kinetics of electrochemical reactions, the higher ionic conductivity through the Grotthuss mechanism in the range of 0– 100  C at a higher temperature is likely the reason for the performance improvements [53]. Though the hydrogen bonding strength decreases with increased temperature, the ionic conductivity by ionic hopping conduction reportedly improves in this temperature range [53]. In order to elucidate the enhanced ionic conduction and electron transfer kinetics with increased temperature, electrochemical impedance spectroscopy was conducted on the AEMWE at the different operational temperatures. As the Nyquist plot shows, both ohmic resistance and polarization resistance decrease consistently as temperature increases (Fig. 6(b), Figure S7). For example, when the cell temperature increases from 50  C to 90  C, reduction ratios are 22.6% (0.239 V cm2 → 0.185 V cm2) for the ohmic resistance and 21.9% (0.219 V cm2 → 0.171 V cm2) for the polarization resistance. The current obtained at 90  C, averaged 0.983 A cm 2 at 1.8 Vcell, is approximately twice of that at 50  C and is one of the high performance among the reported values of the AEMWE (Figures S8 and S9). In summary, we examined the effect of cell operation variables on solid-state alkaline water electrolysis. Pre-feeding the electrolyte 24 h before cell operation accelerates ionic transport through membrane activation. For pressing of MEA, the negative effects due to the change in the structure of the catalyst layer is greater than the reduction in contact resistance from pressing. Since cell assembly torque influences the pressure of contact between the electrodes and membrane, cell integration, gas leakage, chemical mass transport, and mechanical deformation of the electrode, an optimum value of 4 Nm was found in this study. Lastly, the higher cell temperature significantly activates both ionic conduction and catalyst activity so to enhance the water electrolysis performance. This study provides insights to reduce the gap between the developed material properties and device performance in electrolysis, showing meaningful performance enhancement by solely controlling the electrode fabrication and operation factors when utilizing standard materials. Conclusions In this study, electrode fabrication and operation factors including (1) the pressing process during electrode-electrolyte membrane assembly, (2) the cell assembly torque, (3) the electrolyte solution pre-feeding method, and (4) the operating temperature were studied to determine the effects on water electrolysis. The best performance of AEMWE with FAA-3 membranes was approximately 1 A cm 2 at 1.8 V (average of 1 h operation) when the optimum fabrication methods, i.e. no MEA press and 4 Nm cell assembly torque, and operational factors, i.e. 24 h of electrolyte pre-feed and 90  C operation temperature, were applied. The engineering factors affect the electrolysis performance in numerous ways, and ion conductivity in the membrane and catalyst layers, gas permeability, and catalytic activity are largely determined by the fabrication and operation variables. Herein, the crucial effect of external parameters on the performance of AEMWE was revealed by physical and electrochemical

417

analysis on the cell, leading to a current density more than three times higher at 1.8 V following parameter optimization. Although this study used standard noble catalysts, the principles presented in this research can also be applied to AEMWE using non-noble catalysts to achieve high performance, which contributes to commercialization of AEMWE. Acknowledgements This study was supported by the Korean Government through the New and Renewable Energy Core Technology Program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) funded by MOTIE (20153010041750), the Korea CCS R&D Center (KCRC) grant funded by MSIP (No. 2014M1A8A1049349). This study was also financially supported by the KIST through the Institutional Project (2E29600, 2V06840). Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j. jiec.2019.04.007. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]

E. Kintisch, Science 324 (2009) 323. B. Obama, Science 355 (2017) 126. R.A. KERR, Science 222 (1983) 1107. P.P. Edwards, V.L. Kuznetsov, W.I.F. David, Phil, Trans. R. Soc. A 365 (2007) 1043. S. Dutta, J. Ind. Eng. Chem. 20 (2014) 1148. J. Garche, L. Jörissen, Electrochem. Soc. Interface Summer 24 (2015) 39. J.D. Holladay, J. Hu, D.L. King, Y. Wang, Catalysis Today 139 (2009) 244. M.F. Demirbas, Energy Sources 28 (2006) 245. M. Götz, J. Lefebvre, F. Mörs, A.M. Koch, F. Graf, S. Bajohr, R. Reimert, T. Kolb, Renewable Energy 85 (2016) 1371. D. Pletcher, X. Li, Int. J. Hydrogen Energy 36 (2011) 15089. Y. Leng, G. Chen, A.J. Mendoza, T.B. Tighe, M.A. Hickner, C.-Y. Wang, J. Am, Chem. Soc 134 (2012) 9054. U. Babic, M. Suermann, F.N.B. uchi, L. Gubler, T.J. Schmidt, Journal of The Electrochemical Society 164 (2017) F387. M. Carmo, D.L. Fritz, J.r. Mergel, D. Stolten, Int. J. Hydrogen Energy 38 (2013) 4901. M.K. Cho, H.-Y. Park, S. Choe, S.J. Yoo, J.Y. Kim, H.-J. Kim, D. Henkensmeier, S.Y. Lee, Y.-E. Sung, H.S. Park, J.H. Jang, J. Power Sources 347 (2017) 283. H. Osgood, S.V. Devaguptapu, H. Xu, J. Cho, G. Wu, Nano Today 11 (2016) 601. M. Gong, D.-Y. Wang, C.-C. Chen, B.-J. Hwang, H. Dai, Nano Res. 9 (2016) 28. P.T. Babar, A.C. Lokhande, E. Jo, B.S. Pawar, M.G. Gang, S.M. Pawar, J.H. Kim, J. Ind. Eng. Chem. 70 (2019) 116. S. Jung, C.C.L. McCrory, I.M. Ferrer, J.C. Peters, T.F. Jaramillo, J. Mater. Chem. A 4 (2016) 3068. R. Frydendal, E.A. Paoli, B.P. Knudsen, B. Wickman, P. Malacrida, I.E.L. Stephens, I. Chorkendorff, ChemElectroChem 1 (2014) 2075. P.T. Babar, A.C. Lokhande, M.G. Gang, B.S. Pawar, S.M. Pawar, J.H. Kim, J. Ind. Eng. Chem. 60 (2018) 493. W.S. Choi, M.J. Jang, Y.S. Park, K.H. Lee, J.Y. Lee, M.H. Seo, S.M. Choi, ACS Appl. Mater. Interfaces 10 (2018) 38663. J. Lee, H. Lee, B. Lim, J. Ind. Eng. Chem. 58 (2018) 100. H. Kim, J. Kim, S.H. Ahn, J. Ind. Eng. Chem. 72 (2019) 273. L. Zeng, T.S. Zhao, Nano Energy 11 (2015) 110. Q. Duan, S. Ge, C.-Y. Wang, J. Power Sources 243 (2013) 773. H. Yanagi, K. Fukuta, ECS Trans. 16 (2008) 257. D. Aili, A.G. Wright, M.R. Kraglund, K. Jankova, S. Holdcroftb, J.O. Jensen, J. Mater. Chem. A 5 (2017) 5055. D. Chen, M.A. Hickner, ACS Appl. Mater. Interfaces 4 (2012) 5775. X. Ren, S.C. Price, A.C. Jackson, N. Pomerantz, F.L. Beyer, ACS Appl. Mater. Interfaces 6 (2014) 13330. A. Konovalova, H. Kim, S. Kim, A. Lim, H.S. Park, M.R. Kraglund, D. Aili, J.H. Jang, H.-J. Kim, D. Henkensmeier, J. Membr. Sci. 564 (2018) 653.  ska-Piron, Y. Lee, A. Lim, H.S. Park, J.H. Jang, H.-J. Kim, J. A. Marinkas, I. Stru zyn Kim, A. Maljusch, O. Conradi, D. Henkensmeier, Polymer 145 (2018) 242. V. Vijayakumar, S.Y. Nam, J. Ind. Eng. Chem. 70 (2019) 70. W. Xu, K. Scott, Int. J. Hydrogen Energy 35 (2010) 12029. J.E. Park, S.Y. Kang, S.-H. Oh, J.K. Kim, M.S. Lim, C.-Y. Ahn, Y.-H. Cho, Y.-E. Sung, Electrochimica Acta 295 (2019) 99. L. An, T.S. Zhao, Z.H. Chai, P. Tan, L. Zeng, Int. J. Hydrogen Energy 39 (2014) 19869. C.C. Pavel, F. Cecconi, C. Emiliani, S. Santiccioli, A. Scaffidi, S. Catanorchi, M. Comotti, Angew. Chem. Int. Ed. 53 (2014) 1378.

418

A. Lim et al. / Journal of Industrial and Engineering Chemistry 76 (2019) 410–418

[37] H. Ito, N. Kawaguchi, S. Someya, T. Munakata, N. Miyazaki, M. Ishida, A. Nakano, Int. J. Hydrogen Energy 43 (2018) 17030. [38] C. Chaiburi, V. Hacker, Energy Procedia 138 (2017) 229. [39] G. Jeong, M. Kim, J. Han, H.-J. Kim, Y.-G. Shul, E. Cho, J. Power Sources 323 (2016) 142. [40] T. Pandiarajan, L.J. Berchmans, S. Ravichandran, RSC Adv. 5 (2015) 34100. [41] S. Seetharaman, R. Balaji, K. Ramya, K.S. Dhathathreyan, M. Velan, Int. J. Hydrogen Energy 38 (2013) 14934. [42] L. Xiao, S. Zhang, J. Pan, C. Yang, M. He, L. Zhuang, J. Lu, Energy Environ. Sci. 5 (2012) 7869. [43] X. Wu, K. Scott, Int. J. Hydrogen Energy 38 (2013) 3123. [44] X. Wu, K. Scott, J. Power Sources 214 (2012) 124. [45] M. Faraj, M. Boccia, H. Miller, F. Martini, S. Borsacchi, M. Geppi, A. Pucci, Int. J. Hydrogen Energy 37 (2012) 14992.

[46] J. Parrondo, C.G. Arges, M. Niedzwiecki, E.B. Anderson, K.E. Ayers, V. Ramani, RSC Adv. 4 (2014) 9875. [47] J. Parrondo, M. George, C. Capuano, K.E. Ayers, V. Ramani, J. Mater. Chem. A 3 (2015) 10819. [48] C. Chen, Y.-L.S. Tse, G.E. Lindberg, C. Knight, G.A. Voth, J. Am. Chem. Soc. 138 (2016) 991. [49] S. Maurya, S.-H. Shin, Y. Kim, S.-H. Moon, RSC Adv. 5 (2015) 37206. [50] H. Luo, G. Vaivars, B. Agboola, S. Mu, M. Mathe, Solid State Ionics 208 (2012) 52. [51] V.K. Puthiyapura, S. Pasupathi, H. Su, X. Liu, B. Pollet, K. Scott, Int. J. Hydrogen Energy 39 (2014) 1905. [52] A.J. deBethune, T.S. Licht, N. Swendeman, J. Electrochem. Soc. 106 (1959) 616. [53] E. Gileadi, E. Kirowa-Eisner, Electrochem. Acta. 51 (2006) 6003.